oWeatherReader

Weather monitoring system with RTL-SDR and AI-powered recommendations.

Features

• RTL-SDR Weather Monitoring: Monitors weather sensors using RTL-433
• Web API: RESTful API for accessing weather data
• Database Storage: SQLite database for weather reports and recommendations
• AI-Powered Recommendations: Uses Ollama to provide smart home automation recommendations

New Feature: Ollama Integration

The system now queries an Ollama server every 15 minutes (configurable) to get AI-powered recommendations for:

• Whether to operate the air conditioner
• What temperature to set the air conditioner to
• Whether windows should be open
• Weather description with reasoning

Configuration

Configuration can be managed in two ways:

1. Static Configuration: Edit config.json file (loaded at startup)
2. Dynamic Configuration: Use the API endpoints to update configuration at runtime

Configuration schema:

{
  "ollamaServerURL": "http://localhost:11434",
  "ollamaModel": "llama3.2",
  "indoorDeviceModel": "LaCrosse-TX141W",
  "outdoorDeviceModel": "LaCrosse-TX141W", 
  "recommendationIntervalMinutes": 15
}

Dynamic Configuration Updates

You can now update the configuration without restarting the service:

# Get current configuration
curl http://localhost:6656/config

# Update configuration
curl -X POST http://localhost:6656/config \
  -H "Content-Type: application/json" \
  -d '{
    "ollamaServerURL": "http://localhost:11434",
    "ollamaModel": "llama3.2:1b",
    "indoorDeviceModel": "New-Indoor-Sensor",
    "outdoorDeviceModel": "New-Outdoor-Sensor",
    "recommendationIntervalMinutes": 30
  }'

Features of Dynamic Updates:

• Changes take effect immediately
• Configuration is automatically saved to config.json
• If recommendation interval changes, the worker is automatically restarted
• Validation ensures all required fields are provided

API Endpoints

• GET /reports/latest - Get the latest weather report
• GET /reports/:model - Get weather reports for a specific device model
• GET /models - Get all device models with report counts
• GET /recommendations/latest - Get the latest AI recommendation
• GET /config - NEW Get current configuration
• POST /config - NEW Update configuration dynamically

Ollama Response Format

The system expects Ollama to return JSON in this format:

{
  "shouldOperateAirConditioner": true,
  "temperatureToSetAirConditionerInF": 72,
  "shouldWindowBeOpen": false,
  "weatherDescription": "Indoor temperature is 75°F while outdoor is 85°F. AC recommended to maintain comfort."
}

Setup Requirements

1. Ollama Server: Ensure Ollama is running on the configured URL
2. Model: The specified model should be available in Ollama
3. Device Models: Configure the correct device model names for indoor/outdoor sensors

Database Schema

The new OllamaRecommendation table stores:

• Timestamp of recommendation
• AI recommendations (AC operation, temperature, window status)
• Weather description with reasoning
• Indoor and outdoor temperatures used for the recommendation

Building and Running

go build -o oWeatherReader main.go
./oWeatherReader

The service runs on port 6656 by default.